Visual Quality and Accommodation With Novel Optical Designs for Myopia Control

Abstract

Purpose: We evaluated through-focus visual performance and accommodative response in young subjects through three segmented multifocal designs for myopia control, mapped on the spatial light modulator of a monocular adaptive optics visual simulator (AOVS), and compared with single vision (SV).

Methods: The segmented multifocal patterns included a 4 mm diameter center distance zone and offset peripheral defocus (MP1), astigmatism and coma (MP2), or a combination (MP3). High-contrast logMAR visual acuity (VA) was measured with monochromatic stimuli (555 nm). Ocular aberrations were measured using the Hartmann-Shack aberrometry channel. Measurements were taken for distance viewing and five accommodative demands (AD, up to 4.5 D). Accommodative lag was calculated from the dioptric shift of the maximum retinal image quality metric from the corresponding wave aberrations.

Results: Best-corrected logMAR VA was −0.11 ± 0.02 (SV) and slightly reduced by multifocal patterns (−0.08 ± 0.03 [MP1], −0.07 ± 0.04 [MP2], −0.05 ± 0.04 [MP3]). Accommodative lag with SV was lower in emmetropes than myopes (by 0.43D for the largest demand). MP1 significantly decreased accommodative lag in myopes (P = 0.03), unlike MP2 or MP3. Multifocal patterns reduced pupil diameter in myopes at all distances. MP1 improved accommodative response in myopes without compromising distance vision.

Conclusions: AOVS helped to understand the interplay of physiological and lens design factors, potentially guiding custom corrections. A center distance with off-centered positive power in the lens periphery could feature suitable properties (peripheral focus and accommodative focus control) for myopia control.

Translational Relevance: We demonstrate a two-zone contact lens design that provides excellent visual quality and accommodative response, important properties for myopia control lenses.

Introduction

Myopia,¹ a growing public health concern, is anticipated to affect half of the global population by 2050, highlighting the urgent need for effective interventions.² This condition not only adversely impacts vision-related quality of life, but also poses risks of pathological complications, such as choroidal neovascularization and myopic macular degeneration.^3–5 

In response to this critical issue, extensive research efforts have been dedicated to understanding the factors influencing myopia progression and developing innovative strategies to slow down its progression.^6–8 One of the strategies gaining traction is the use of multifocal contact lenses (MCLs).⁹ Clinical trials have shown promising outcomes. Some MCL designs, in particular center-distance lenses and relatively high add powers, appear to have been the most successful in reducing myopia progression^10–12 although not universally. The most recurrently proposed working principle for center-distance MCLs for myopia control revolves around the induction of relative myopic defocus in the peripheral retina, which would act as a protective measure against foveal axial growth.^13–15 However, the effectiveness of MCLs in avoiding hyperopic defocus and inducing myopic defocus in young eyes in the periphery, but also the central retina, is highly contingent on the accommodative response of the individual eye when wearing these lenses. More precisely if the eye relaxes its accommodation to utilize the near add for tasks requiring near vision, the expected myopic defocus may not be realized,¹⁶ because the eye would be subject to the hyperopic blur associated with the out-of-focus distant image. 

There are numerous reports in the literature indicating that myopes exhibit higher accommodative lags than emmetropes.^17–21 Controversy arises on the magnitude or even the presence of the effect and the suitability of techniques such as standard autorefractometry to accurately measure the residual refractive error.²² However, studies that use more refined methods (i.e., aberrometry and retina image quality metrics) confirm the presence of a larger residual hyperopic defocus in accommodating myopes²³ compared to emmetropes, including a recent a study from our group,²⁴ using similar methods to those used in the current study. 

In addition to studies of accommodation under natural viewing conditions, several investigations have examined the accommodative response in the presence of altered aberrations or with myopia control corrections. For example, inducing positive spherical aberration (such as in orthokeratology²⁵) or positive and negative spherical aberration and coma in a deformable mirror of an adaptive optics (AO) visual simulator has been shown to impact the accommodative response.^26–28 The interplay of a patient's aberrations, lens design, and pupil diameter has been recognized as critical in determining the accommodative response.¹⁶ On the other hand, objective measurements through MCLs on the eye may be challenging, particularly with zonal or diffractive designs. An alternative to investigate the effect of different lens designs on vision and the accommodative response is the use of AO technologies. Generally, segmented diffractive multifocal corrections are mapped onto a spatial light modulator (SLM), whereas smooth profiles, for example inducing spherical aberration, are well represented on a deformable mirror (DM). 

Decoupling the tested lens (which is mapped either in the DM or the SLM) from the eye allows accurate measurement of the wave aberrations of the eye alone at different accommodative demands. On the other hand, the combined effect of the eye's aberrations and the lens design is captured by adding the wave aberration and the phase map representing the lens, allowing the calculation of the accommodative response (and lag) through the lens. With this paradigm, Vedhakrishnan et al.¹⁶ estimated the accommodative response in young adults while viewing a target through various zonal center-near and center distance designs mapped on an SLM or spherical aberration mapped on a DM. That study showed that the lowest accommodative lag occurred with a center distance design, followed by positive spherical aberration. 

Besides the passive shift in best focus produced by spherical aberration, or by patterns with a certain spatial distribution in power (for example zonal concentric), the dependency of the accommodative response on lens design is intriguing and has stimulated several investigations on what really prompts subjects to accommodate better through certain designs. A recent publication by Gifford et al.²⁹ underscores that it is the multifocal lens design that shapes accommodative response. The authors investigated the effect of four different types of commercially available MCL on accommodation and found that young myopes wearing aspheric MCL experienced a reduction in their accommodation lag, regardless of the specific “add” power used. Among the different lenses tested, the concentric dual-focus lens produced smaller lag of accommodation, but it also demonstrated greater variability in the accommodative response compared to the other lenses. In another study, large transition zones emerged as a major factor influencing accommodation, particularly to intermediate distance.³⁰ Wagner et al.³¹ showed that subjects can be trained using biofeedback to improve the accommodative response with MCLs, suggesting that accurate accommodative response in turn will enhance the myopia control properties of the lens. In general, this prior evidence highlights the need to tailor the design and correction for increased accommodation accuracy.^32,33 

Despite the potential benefits of MCLs, challenges persist, including limited systematic testing of lens design parameters due to fixed designs of commercially available lenses. AO visual simulators emerge as an ideal experimental platform for exploring potential variations in the accommodative response among different multifocal lens designs. First, AO facilitates noninvasive simulation of lens designs without the need to physically place the contact lens on the eye. Additionally, because these lens designs are digitally programmed, they are not constrained by commercial or physical limitations.^6,16,34 From a clinical standpoint, AO visual simulators offer the opportunity to evaluate vision in individuals with various MCL designs before the actual lens fitting and test both visual performance with these lenses because preserving excellent vision both at distance and near is key and the accommodative response. This process eventually facilitates the selection of the lens that optimizes perceived visual quality and overall visual performance.³⁵ Furthermore, an integral component of the AO system, the Hartmann-Shack (HS) wavefront sensor, proves valuable in quantifying the accommodation response. This involves measurements of low- and high-order aberrations, estimates of retinal image quality, and calculations of the focus shift required for the best image quality on the retina.¹⁶ 

In this study we assessed the effect of different segmented multifocal designs on the optical quality and the accommodative responses of young adult emmetropes and myopes. A companion study evaluated baseline differences in through focus visual acuity (TFVA) and accommodative lag between myopes and emmetropes with no multifocal corrections.²⁴ Specifically, the aim of the current study was to test visual performance and accommodation through new segmented lenses that were designed to modulate peripheral image quality. The peripheral retina experiences anisotropic blur primarily caused by oblique astigmatism, known as "odd-error" blur signals.³⁶ The orientation of the peripheral blur (radial or circumferential) may then serve as an optical cue for the eye to determine the sign of defocus.^37,38 We hypothesize that the optical manipulations produced by the proposed designs may play a role in accommodation and emmetropization mechanisms. In a prior study, Zheleznyak³⁷ investigated the nature of peripheral blur in different subject groups. The study revealed that all subjects exhibited anisotropic blur in the periphery, suggesting varying degrees of elongation in different orientations. Specifically, myopes demonstrated a circumferentially elongated peripheral blur, whereas emmetropes and hyperopes exhibited radial blur. Built on this principle, the designs used in the study were purposely designed to modify the optical anisotropy in the peripheral retina of the average myopic eye. Design 1 intends to rotate the blur orientation from circumferential to radial in the periphery of the average myopic eye by inducing defocus to interact with the eye's native peripheral wavefront aberrations. Design 2 intends to correct the average myopic eye's peripheral wavefront aberrations, thereby removing optical anisotropy cues and improving peripheral image quality. Design 3, like Design 2, corrects the peripheral aberrations and, in addition, induces positive defocus to place a circular myopic blur circle in the peripheral retina of the average myopic eye. Although these optical designs were formulated with peripheral retinal image quality in mind, the current study only stimulated the central fovea. 

The current study investigated through-focus visual performance with these designs, and their potential impact of altering accommodative lag. Given prior evidence, we may anticipate that efficacy as a myopia control treatment may be achieved by the lenses that stimulate a better accommodative response. The findings using AO will guide final designs for a more extensive testing in a controlled clinical trial with real contact lenses. 

Methods

TFVA and accommodative response were measured for three multifocal patterns consisting of a central plano optical zone, and a peripheral optical zone segmented into four identical quadrants containing peripheral aberrations and a single vision control pattern. The lenses were simulated in the SLM of an AO Visual Simulator. TFVA was measured psychophysically under natural accommodation and cycloplegia with fixed pupil diameter (5 mm pupil). The eye's aberrations, residual defocus and pupil diameter were measured as a function of accommodative demand while the subject accommodated to a monochromatic stimulus. Retinal image quality metrics, calculated from the measured wavefront data, were used to estimate the accommodative lag. 

Subjects

A total of 20 subjects participated in the study, with ages ranging from 24 to 29 years (25.4 ± 1.1 years). The subjects were divided into two refractive groups (n = 10 per group): (1) Myopes (Spherical error [SE]: −2.00 to −6.25 D; (2) Emmetropes (SE: 0.25 to 0.75 D). Astigmatism in all cases was below −0.50 D. The study protocols met the tenets of the Declaration of Helsinki and had been approved by the University of Rochester Institutional Review Board. All participants were informed about the study and experimental procedures and signed informed consent forms prior to any study procedures. 

Segmented MCLs

Three novel segmented center-distance multifocal patterns proposed by Clerio Vision (Rochester, NY, USA) were tested. The multifocal patterns (MPs) consisted of a central 4 mm diameter plano optical zone for distance, and four peripheral segmented regions (Fig. 1, shown as gray scale phase maps, using a 2π-wrapping). Prior research³⁸ had assessed the optical quality and anisotropy of myopic eyes across the visual field, using population-averaged data on wavefront aberrations as a foundation.³⁹ The three designs were crafted with the specific goal of altering the peripheral blur, particularly at a 30 deg angle, all while preserving the clarity of central vision, with a center-distance 4 mm zone and 7.5 mm full diameter. Figure 2 illustrates for an average myopic eye the simulated through-focus (TF) optical quality (defined as the volume under the Modulation Transfer Function, MTF, Fig. 2A) and optical anisotropy (Fig. 2B) at 30° nasal field, respectively, for the three lenses and a plano control. 

Figure 1.

 

Phase maps of the MP simulated in the SLM with a central plano optical zone and a peripheral optical zone segmented into four identical quadrants containing peripheral aberrations, off-centered by 2.6 mm in the horizontal and vertical meridians. MP1 contains pure defocus. MP2 contains astigmatism and coma. MP3 contains a combination of defocus, astigmatism, and coma in the periphery. The grayscale color bar represents the optical path length from 0.0 (black) to 1.0 (white) waves at 555 nm.

Figure 2.

 

(A) Optical quality and (B) optical anisotropy for the average myopic eye at 30° in the nasal visual field³⁷ for the four optical conditions investigated in this study: a plano control, MP1, MP2, and MP3.

The first pattern (MP1, Fig. 1 left) contains pure defocus (0.75 µm, equivalent to 1.30 D for a 4 mm pupil, 2.6 mm off-center). By inducing pure defocus, the curves in Figure 2 are shifted horizontally. This allows the minimum in the through-focus optical anisotropy (Fig. 2B) to occur on the retina, rather than in front or behind. Said another way, MP1 intends to change the orientation of the peripheral blur from being circumferentially elongated to radially elongated. 

The second pattern (MP2, Fig. 2 middle) contains astigmatism and coma (0.1 µm defocus, −0.82 µm astigmatism and 0.25 µm coma, 2.6 mm off-center). This design is intended to correct peripheral astigmatism and coma in the average eye at 30°.³⁹ By correcting the peripheral aberrations, peak image quality is increased, and optical anisotropy is near unity, as shown in Figure 2. 

The third pattern (MP3, Fig. 1 right) contains a combination of defocus, astigmatism, and coma (0.75 µm defocus, −0.82 µm astigmatism and +0.25 µm coma, 2.6 mm off-center). This design is a combination of MP1 and MP2, neutralizing astigmatism and coma while adding 1.3 D of myopic defocus.^40–42 This brings best focus in front of the retina, and like MP2, has through-focus optical anisotropy near unity, as seen in Figure 2. 

The phase maps illustrated in Figure 1 correspond to patterns as mapped onto a reflective, phase-only SLM. The wavefront, calculated in Matlab software (MathWorks Inc., Natick, MA, USA), was represented in the SLM as a grayscale image phase-wrapped at 1.0 waves (555 nm wavelength). 

Experimental Set-Up

The experiment was performed using a custom-developed AO system at the Center for Visual Science (University of Rochester, Rochester, NY, USA), which represents an evolution of a prior design developed at the VioBioLab (Institute of Optics, CSIC, Madrid, Spain), which is described in detail in prior publications.^35,42–50 The system illumination is provided by a supercontinuum laser source (Rock 450 4; Leukos, Limoges, France) coupled with acoustic-optic tunable filters (Optics and Photonics Gooch & Housego; G&H, Ilminster Somerset, UK) to select the desired wavelength: visible 555 nm (5 nm bandwidth) to illuminate the visual stimulus, and infrared wavelength 830 nm for aberration measurements. The Hartmann-Shack wavefront sensor (32 × 32 microlenses; HASO 32 OEM; Imagine Eyes, Orsay, France) was used to measure the eye's aberrations, from which the accommodative response is calculated (Aberration Measurements). An electromagnetic deformable mirror (52 actuators and 50 mm stroke; MIRAO; Imagine Eyes) was used in this study only to correct the optical system's aberrations. A calibrated tunable lens (EL-10-30-TC, with negative achromatic lens offset lens; Optotune Switzerland AG, Dietikon, Switzerland) was used to correct the subject's refractive error and to induce defocus (accommodative demands). A pupil monitoring system (LED ring illuminator and a CCD camera) was used to align and monitor eye centration. A reflective phase-only SLM (LCoS-SLM, 1920 × 1080 pixels; Holoeye, Berlin, Germany) conjugate to the subject's pupil was used to simulate the lens patterns. The visual stimulus in the psychophysical channel was displayed on a digital micro-mirror device (DMD) with an angular subtend of 2°. An artificial pupil was placed in a conjugate pupil plane in the visual stimulus channel and was fully open for experiments under natural viewing conditions or limited to 5 mm pupils in experiments under cycloplegia. All optoelectronic elements are controlled with two computers by commercial software and custom C++ and Matlab software. 

Experimental Procedure

Each subject was measured in two different sessions. Session 1 involved measuring TFVA, and the eye's wave aberrations under natural conditions (i.e., natural pupil and free accommodation) for a series of accommodative demands (0 to 4.5 D), while targets were viewed through the multifocal patterns MP1, MP2 and MP3, and a single vision (SV) control condition. Session 2 involved performing TFVA measurements under cycloplegia (1% tropicamide ophthalmic solution, USP Alcon Laboratories, Inc. Fort Worth, TX, USA; two drops before the experiment and repeated every hour) and 5 mm fixed artificial pupil, through MP1, MP2, MP3, and SV. Measurements under natural accommodation allow to assess quality of vision under natural pupil dynamics, and accounting for the effect of accommodation, also including the effects of the natural aberrations of the eye, as well as the coupling of those with the multifocal patterns both at far and near. Measurements under paralyzed accommodation and a fixed pupil diameter, allow direct comparisons of lens performance across individuals, and account for the effect of the coupling of the design and the aberrations of the eye, excluding the effect of accommodation. 

Visual stimuli (E letter for VA measurements and Maltese cross for accommodation and focus adjustment) were illuminated with monochromatic light (555 nm) and viewed monocularly (the contralateral eye was covered with a patch). Subjects were aligned to the system and adjusted the best focus at distance (starting with positive defocus) using a keypad to change the power of the tunable lens. The zero-defocus was obtained from the average of at least 5 focus settings repetitions, for each pattern, for both natural accommodation and paralyzed accommodation, and it varied by less than 0.50 D across conditions. The duration of Session 1 was 3.5 hours, with frequent breaks. The duration of Session 2 was approximately 1.5 hours. All measurements were performed monocularly in a darkened room. Before starting the measurements, subjects were instructed on the nature of the experiment and trained with several practice trials. 

TFVA

VA was assessed using a tumbling high contrast E letter test in 4-Alternative Forced Choice procedure with a QUEST algorithm (from Matlab Psychtoolbox package).⁵¹ The black E letter was displayed for 0.5 seconds on a green (555 nm) background. Participants indicated the orientation of the E letter, which changed size according to the QUEST. Each VA measurement consisted of 40 trials, with a threshold criterion of 75%. The threshold VA measurement was calculated by averaging the last 10 stimulus values. VA is reported in units of logMAR. The measurements were conducted within a range of 1 D to −3 D, in 0.5 D increments around the best focus (0 D). Measurements were done for all multifocal patterns and SV, under natural and paralyzed accommodation. 

Aberration Measurements

Aberrations were measured in IR (830 nm) while the subject viewed a monochromatic (555 nm) Maltese cross stimulus. HS images from the subject's eye were collected in IR through the SLM. Wave aberrations were calculated from the Hartmann-Shack images and fitted to seventh order Zernike polynomials following OSA standards,^26,52 and presented as the average of five repeated measurements. Natural pupil diameters were also obtained from the HS images acquired in Session 1 (and the aberrations estimated for the corresponding pupil diameter) for all conditions (SV, MP1, MP2, and MP3) and accommodative demands. Pupil diameters were set to 5 mm for images acquired in Session 2. 

A baseline measurement of potential contributions of the SLM in IR for each pattern was obtained for a diffraction-limited artificial eye (with the system and artificial eye's aberrations corrected by the deformable mirror) and subtracted from the test images from the subject's eye. Control experiments to verify the accuracy of this method were performed mapping known aberrations in the deformable mirror (known amount of spherical aberration, astigmatism at 45° and horizontal coma) and measuring them with the HS wavefront sensor. The induced aberrations were measured within 98% accuracy. 

Accommodative Lag Estimation

Figure 3A illustrates (for one subject and MP1) the estimation of the accommodative lag as a function of accommodative demand through the multifocal patterns, from the wave aberrations. The corresponding eye's wave aberration map is added to the phase map representing the multifocal pattern, taking into account the eye's pupil diameter. The upper row represents the eye's aberrations (best focus at 0 D and corresponding Zernike defocus for each accommodative demand 0–4.5 D), revealing increase in defocus, increase in negative spherical aberration and decrease pupil diameter with increased accommodation. The middle row represents the multifocal pattern as mapped on the SLM, cropped to the corresponding pupil diameter. The lower row shows the addition of the eye's wave aberration phase and the multifocal pattern phase map. 

Figure 3.

 

(A) Calculated wave aberrations of eye (measured, upper row), multifocal pattern (middle row) and added wave aberration and multifocal phase map (lower row). Data are for different accommodative demands (0–4.5 D) and the data below represent the corresponding pupil diameter (mm). The example corresponds to Subject 9 (age = 25 years; spherical error = −2 D) and multifocal pattern MP1. (B) Through-focus volume under the MTF in 3–5 c/° range (MTF3–5) for single vision (SV, left, gray curves) and MP1 (right, green curves), for all accommodative demands (0–4.5 D) Accommodative lags are estimated from the displacement of the peak in each curve from zero defocus.

The residual defocus for each accommodative demand (0–4.5 D) was calculated from the wave aberrations, using Fourier Optics, as described previously.^16,53 For every wave, aberration the volume under the MTF in the 3–5 c/° (MTF3-5)^53,54 was calculated TF. The defocus shift (relative to 0 D, which maximizes quality at far in the distance viewing measurements) for each accommodative demand is taken as an estimate of the residual refractive error (accommodative lag). As an illustration, Figure 3B shows the calculated TF MTF3-5 for SV (left) and MP1 (right) at all accommodative demands. A lateral offset of the peak toward positive values indicates hyperopic defocus. 

Data Analysis

TFVA curves (natural and paralyzed accommodation), pupil diameters, and accommodative lag (natural conditions) were compared across lenses (SV, MP1–MP3) in the myopic and emmetropic groups. The following metrics were quantified from the data: (1) Absolute VA at Far; (2) Degradation at far, calculated as the difference in logMAR VA at far between SV and a given MP; (3) Improvement at near, calculated as the difference in logMAR VA at near (−2.50 D) between SV and a given MP; (4) Visual imbalance, defined as the standard deviation of logMAR VA, in the 0 D to −3 D range); (5) Depth-of-Focus (DOF), defined as the dioptric range (from 0 to near) over which VA is 0.2 logMAR or better.⁵⁵ When the range exceeded 3 D (the entire range measured at near) it was accounted as 3 D; (6) Absolute pupil diameters, and slope of the change in pupil diameter with accommodative demand; (7) Slope of the accommodative lag. 

Statistical analysis was performed using Matlab. The normality assumption was assessed using the Shapiro-Wilk test. To compare across subjects and different accommodation states specific non-parametric tests were used. The main non-parametric test used were: (1) Mann-Whitney U test for comparing the difference between myopes and emmetropes under different conditions; (2) Wilcoxon signed-rank test for comparing natural and paralyzed accommodation within the same group; (3) Kruskal-Wallis test for comparing different conditions across lenses to identify significant differences across SV and MP conditions. 

Results

Visual Performance With Multifocal Patterns

Figure 4 shows representative example in a myopic subject (Subject 4; age = 25 years; spherical error = −2 D) and emmetropic subject (Subject 19; age = 26 years, spherical error = 0.5 D) of TFVA curves for natural viewing (Fig. 4A, upper panels) and paralyzed accommodation (Fig. 4B, lower panels), through the three different multifocal patterns (MP1, green; MP2, blue; and MP3, red) and SV (gray). Qualitatively, VA at far is only reduced with respect of other conditions in the myopic eye for MP3 under natural accommodation. In the SV condition, natural accommodation improves vision at near in the emmetropic but not in the myopic eye, with respect to the paralyzed accommodation/5 mm pupil condition. In the myopic eye, the multifocal patterns, to a larger extent MP1, increase visual quality at near. The improvement at near is most likely associated to a better accommodative response, as it is not found under paralyzed accommodation. In the emmetropic eye, differences in TFVA with natural and paralyzed accommodation are smaller (although tendencies across multifocal patterns go in the same direction), probably because of more comparable pupil diameters between those conditions. 

Figure 4.

 

TFVA in two representative subjects (myopic S 4 and emmetropic S 19). (A) Under natural accommodation (upper panel, dark lines). (B) Under paralyzed accommodation (lower panel, lighter lines).

Figure 5 shows average VA at far across subjects, myopes (Fig. 5A) and emmetropes (Fig. 5B) and conditions. VA varies across subjects (error bars represent standard deviations), between that for natural pupils (average values shown below the bars, in mm) and for the 5 mm pupil diameter, and across MP1, M2, MP3, and SV. VA at far is above 0.0 logMAR in all cases for SV under natural conditions, and only dropped lower below 0.0 logMAR in two myopic (subject S 12 and S 16) and one emmetropic subject (S 6). The analysis revealed no significant difference in VA at far between myopic and emmetropic eyes under natural accommodation (Mann-Whitney U test, P = 0.293). However, significant differences between groups were observed under paralyzed accommodation (Mann-Whitney U test, P = 0.01), specifically for myopic eyes (Wilcoxon signed-rank test, P = 0.04), whereas for emmetropic eyes, the difference was not statistically significant (Wilcoxon signed-rank test, P = 0.45). Additionally, significant differences were found across the different conditions (SV and MP) for both myopic and emmetropic eyes (Kruskal-Wallis test, P < 0.05). 

Figure 5.

 

Average logMAR visual acuity at far across subjects. (A) Myopic group, solid bars. (B) Emmetropic group, patterned bars. The different lenses are represented by different colors: black/gray for SV, green for MP1, blue for MP2, and red for MP3. Natural accommodation is represented by dark bars and paralyzed accommodation by light bars. The numbers below the bars represent the average pupil diameter in each condition, in mm, ranging from 6.9 to 6.3 mm under natural accommodation and fixed 5 mm pupil under paralyzed accommodation.

Figure 6 shows the difference from the SV control at far and the improvement at near for myopes (solid bars) and emmetropes (patterned bars) produced by the multifocal patterns (MP1-green, MP2-blue and MP3-red) at far (best focus) and near (−2.5 D), respectively.³⁷ Positive values in the y-axis represent an improvement, and negative values represent a degradation in visual performance produced by the multifocal pattern tested. 

Figure 6.

 

Average degradation at far (left panels) and improvement at near (right panels), for (A) natural accommodation and (B) paralyzed accommodation. Data from myopes are represented by solid bars and from emmetropes by patterned bars. Data are estimated as differences between LogMAR VA with the multifocal patterns and SV control. Multifocal patterns are identified by color: Green = MP1; Blue = MP2; Red = MP3. Error bars stand for standard deviations across subjects for each condition.

For natural accommodation at far (Fig. 6A, left) on average, the multifocal patterns produced a higher degradation (Fig. 6A, left) in myopes than in emmetropes. In both groups the degradation was lowest for MP1 (−0.02 ± 0.01/−0.03 ± 0.01 logMAR) than for MP2 (−0.03 ± 0.01/−0.04 ± 0.01 logMAR) and MP3 (−0.04 ± 0.01/−0.07 ± 0.01 logMAR), for emmetropes/myopes respectively. The differences in degradation at far between myopic and emmetropic eyes were statistically significant for MP1 ((Mann-Whitney U Test, P = 0.04) and MP3 (P = 0.03) but not MP2 (P = 0.62) 

Under paralyzed accommodation at far (Fig. 6B, left), the degradation was also lowest for MP1 (−0.02 ± 0.03/−0.02 ± 0.02) and MP2 (−0.02 ± 0.02/−0.06 ± 0.05) than MP3 (−0.03 ± 0.04/−0.04 ± 0.03). However, the differences in the degradation at far under paralyzed accommodation between myopes and emmetropes were not statistically significant (Mann-Whitney U test, P = 0.82 for MP1, P = 0.62 for MP2, and P = 0.67 for MP3). 

For natural accommodation at near (Fig. 6A, right), only MP1 showed an improvement compared to SV in emmetropes and myopes (0.03 ± 0.04/0.02 ± 0.01 logMAR), which was statistically significant (Mann-Whitney U test, P = 0.05). The larger standard deviation in emmetropes may be indicative of a higher accommodative performance in emmetropes than in myopes in the SV, resulting in a less systematic improvement of a VA at near with MF in emmetropes. In contrast, neither MP2 (−0.02 ± 0.02/−0.04 ± 0.03 logMAR) nor MP3 (−0.10 ± 0.01/−0.08 ± 0.02 logMAR) produced improvement at near, and in fact degraded performance at near in both emmetropes and myopes (Mann-Whitney U Test, 0.62 for MP2, and 0.57 for MP3). Under paralyzed accommodation at near (Fig. 6B, right) all MPs produced an improvement at near (Fig. 6B, right), with notably the highest improvement obtained for MP1 (0.14 ± 0.04/0.11 ± 0.05 logMAR, Mann-Whitney U test, P = 0.05), compared with MP2 (0.03 ± 0.06/0.02 ± 0.06 logMAR, not significant) and MP3 (0.03 ± 0.07/0.08 ± 0.05 logMAR, Mann-Whitney U test, P = 0.03). 

As expected, there were significant differences in near acuity between natural and paralyzed accommodation for both emmetropic (Wilcoxon signed-rank test, P = 0.002) and myopic eyes (Wilcoxon signed-rank test, P = 0.003). The marked difference in the effect of the multifocal patterns on performance at near between natural and paralyzed accommodation suggests that, under natural accommodation MP1 couples with the optics of the eye and the accommodation system to drive accommodation and improve performance at near, although this was not the case with MP2 and MP3. 

Figure 7 shows the visual imbalance metric,³⁴ accounting for the variations of VA across the TFVA curve from far to near (0 to −3 D), for myopes (Fig. 7A) and emmetropes (Fig. 7B). Visual imbalance was consistently higher under paralyzed accommodation compared to natural aberrations, statistically significantly in emmetropes (Fig. 7A) for all conditions (Mann-Whitney U test, P = 0.01 for SV, P = 0.002 for MP1, P = 0.03 for MP2, and P = 0.01 for MP3), and in myopes (Fig. 7B) in all but one condition (Mann-Whitney U test, P = 0.04 for SV, MP1, and MP3). Specifically, the mean difference in visual imbalance between natural and paralyzed accommodation with SV was 0.11 logMAR for emmetropes and 0.08 logMAR for myopes, and with MP1 was 0.07 logMAR for emmetropes and 0.05 logMAR for myopes. Under natural accommodation the lowest visual imbalance is produced by MP1 (Mann-Whitney U test, P = 0.016), again suggesting a more sustained accommodation with this pattern. 

Figure 7.

 

Visual imbalance for (A) myopic group (solid bars) and (B) emmetropic group (patterned bars). Data for natural accommodation are represented in dark color and paralyzed accommodation in light color. Data are estimated as the standard deviation of logMAR VA in the TFVA curves, from far to near. Different lenses are identified with different colors (Black = SV; Green = MP1, Blue = MP2; Red = MP3). Error bars stand for standard deviations across subjects for each condition.

Figure 8 shows the average DOF obtained from the TF curves as defined in (Data Analysis), for natural (Fig. 8A) and paralyzed accommodation (Fig. 8B) for myopes (solid bars) and emmetropes (patterned bars). DOF was statistically significantly higher in emmetropes than in myopes under natural accommodation, consistent with their more sustained accommodative response, and was highest for MP1 (over SV, MP2, MP3). In the case of paralyzed accommodation (Fig. 8B), no significant increase in DOF was observed with multifocal lenses compared to single vision lenses for either group. 

Figure 8.

 

DOF in diopters (D) for (A) natural accommodation (dark colors) and (B) paralyzed accommodation (light colors). The data are presented for myopes (solid bars) and emmetropes (patterned bars). Error bars represent the standard deviations across subjects for each condition. Different lenses are identified with different color (Black = SV; Green = MP1; Blue = MP2; Red = MP3. Error bars stand for standard deviations across subjects for each condition.

Pupil Diameter With Multifocal Patterns

Figure 9 shows average the pupil diameter in the accommodative demand range from 0 to −4.5 D in myopes and emmetropes. Myopic eyes (Fig. 9A) had significantly larger pupils than emmetropes, at all accommodative demands (Kruskal-Wallis test, P = 0.005). Notably, the presence of multifocal patterns decreased pupil diameter significantly in myopes (Kruskal-Wallis test, P = 0.01), but not in emmetropes (Fig. 9B) (Kruskal-Wallis test, P = 0.062). There were not statistically significant differences in pupil diameters across the different multifocal patterns in either group, or between emmetropes and myopes with multifocal patterns. The variability in repeated measurements of pupil diameter did not change with accommodative demand, and was on average 0.10 mm. The intersubject variability in pupil diameter (standard deviation, averaged across conditions) was higher in myopes than emmetropes (0.36 mm/0.27 mm, respectively). Pupil diameter decreased with accommodation in myopes, with slope values of −0.43, −0.45, −0.41, and −0.44 mm/D (SV, MP1, MP2 and MP3) and for emmetropes as −0.32, −0.32, −0.35, and −0.35 mm/D (SV, MP1, MP2, and MP3). Although the presence of any multifocal pattern decreased the pupil diameter with respect to the natural viewing in myopes, the presence of a multifocal pattern did not have an impact on the slope of the pupil diameter vs accommodative demand function. 

Figure 9.

 

Change in pupil diameters change with accommodative demand in (A) myopes (solid lines) and (B) emmetropes (dashed lines). Data are for single vision control (black lines) and multifocal patterns MP1 (green lines), MP2 (blue lines), and MP3 (red lines). Error bars represent standard deviations across subjects. Accommodative demand is expressed in absolute values.

Accommodative Lag With Multifocal Patterns

Figure 10 shows the accommodative lag (as described in Accommodative Lag Estimation) for each accommodative demand. The top plots (Fig. 10A) compare accommodative lag curves between myopes (solid line) and emmetropes (dashed line). The lower plots (Fig. 10B) compare the average accommodative lag curves for SV and with the multifocal patterns for myopes (left curves) and emmetropes (right curves). Lag was significantly higher in myopes than emmetropes for SV control (Mann-Whitney U test, P = 0.01). The largest accommodative lags were obtained with MP2, followed by MP3, and the lowest lag was obtained with MP1. In comparison with the SV control condition, MP1 was the only multifocal pattern that reduced the accommodative lag significantly in myopes, with a decrease of 0.31 D at −3 D accommodative demand (Wilcoxon signed-rank test, P = 0.03). 

Figure 10.

 

(A) Average accommodative lag in emmetropes (solid lines) and myopes (dashed lines) for the SV control, and the three multifocal patterns. (B) Accommodative lag with single vision control and multifocal patterns (MP1, MP2, and MP3) in myopes (left) and emmetropes (right). Error bars stand for the standard deviation across subjects. Accommodative demand is expressed in absolute values.

Figure 11 shows the slopes of the accommodative lag curves in the 0 to −4.5 D range (left) and of 0 to −3 D range (right), averaged across subjects: myopes (patterned bars) and emmetropes (solid bars). The higher slopes were found for MP2 and MP3, and the lower slopes for MP1 and SV. Also, the lag slopes were consistently lower for emmetropes than for myopes. In the 0 to −3 D range, MP1 produced a statistically significant decrease in slope, which was more pronounced in myopes (Wilcoxon signed-rank test, P = 0.03) than emmetropes (Wilcoxon signed-rank test, P = 0.04). MP2 and MP3 did not produce statistically significant changes in the accommodation lag slopes. 

Figure 11.

 

Average slopes of accommodative lag curves (A) for 4.5D accommodative demand range and (B) for 3D accommodative demand range.

Discussion

The current study evaluated novel optical designs for myopia control contact lenses using an AO vision simulator. The optical designs were formulated to manipulate various aspects of the peripheral retinal image, such as peripheral optical quality, defocus and optical anisotropy (i.e., directionality of blur). While the optical designs were motivated by peripheral optical aberrations, it is crucial to also examine the impact of contact lens designs of aspects of vision besides myopia control efficacy, namely the foveal functions of resolution acuity and accommodative response. We found that all designs provided satisfactory distance VA (better than 20/20) and did not exacerbate significantly accommodative lag for near objects, and in fact one of the designs (MP1) significantly decreased lag. 

Previous work has shown that the peripheral retina is crucial for emmetropization and can be used as a target for optical therapies intended to slow the rate of myopic progression. Because contact lenses sit several millimeters from the eye's pupil, the peripheral optical zone of a contact lens can be customized to address the optical quality in the peripheral retina. The designs herein have a large 4 mm central plano zone, to ensure un-aberrated foveal vision, while the peripheral optical zone is intended for the myopia control therapy. 

Many studies have reported the average wavefront aberrations across the eye's field of view, which are dominated, off-axis, by astigmatism and coma. These circularly asymmetric aberrations consequently produce an asymmetric point spread function, whose size and orientation are affected by defocus. Although the signal for retinal defocus is encoded by the size and orientation of retinal blur, it is unknown whether peripheral optical anisotropy is used by the visual system as a cue for discerning the sign of retinal defocus. The optical designs under investigation were intended to address this question by modulated peripheral optical anisotropy. As described above, MP1 intends to flip the orientation of the peripheral blur in the average myopic eye from circumferential to radial, akin to the peripheral anisotropy of emmetropes and hyperopes,³⁷ and co-aligned with retinal ganglion cell receptive fields.^56–60 Alternatively, designs MP2 and MP3 intend to remove the peripheral anisotropy cue by correcting peripheral astigmatism and coma. Finally, MP2 and MP3 differ, in that MP3 has additional positive defocus to provide the peripheral retina with a myopic defocus. A separate study⁶¹ is underway evaluating the efficacy of these contact lens designs. Preliminary results monitoring short term changes in axial length, a biomarker of myopia control, showed that MP1, MP2 and MP3 led to a reduction in axial length (−13 ± 6, −4 ± 8, −6 ± 9 µm) after 30 minutes of viewing, indicating a future deceleration of eye growth. By comparison, a dual-focus multifocal design and a plano control resulted in −7 ± 10 and −2 ± 5 µm of axial length decrease. 

Therapies for diseases are typically evaluated along two major axes: safety and efficacy. At present, one contact lens has been FDA approved for myopia control (MiSight; CooperVision, San Ramon, CA, USA), and clinical studies have assessed its efficacy¹¹. In a recent study, we also evaluated visual performance and accommodation with that lens both centrally and peripherally.⁶² The current study is analogous to a safety assessment of a therapeutic contact lens to combat myopia progression, where safety is defined as visual performance. As such, we evaluated both high contrast VA and accommodative response at viewing distances ranging from far to near with the novel designs, simulated in an AO visual simulator⁴¹ (i.e., before placing the contact lenses on eye). The results showed that for both emmetropes and myopes, VA at far distances exhibited negligible differences across the test conditions. These findings suggest that both groups maintained a stable VA across the different conditions, indicating minimal variation in distance vision under these tested patterns. 

The lag of accommodation, however, demonstrated more variability between the two groups and across the different conditions. For emmetropes, the lag ranged from 0.34 diopters (MP1) to 0.72 diopters (MP2), with the highest lag observed during the MP2 condition. In contrast, myopes exhibited a lag ranging from 0.35 D (MP1) to 0.82 D (MP3), with the highest lag measured with MP3. Notably, MP1 was the only pattern that reduced the lag of accommodation for both emmetropes and myopes. This suggests that the MP1 pattern is particularly effective in enhancing accommodation without compromising distance VA, making it a promising candidate for improving visual performance in myopia. 

The observed minimal degradation in VA can be expected due to the large un-aberrated central optical zone of all designs. Because pediatric pupil sizes vary widely (2 to 8 mm in diameter),^63,64 the trade-off in surface area of the (a) central optical zone for high quality central vision versus the (b) peripheral optical zones for delivering myopia control aberrations led us to the compromise of a 4 mm diameter central plano optical zone. 

Currently available contact lens-based myopia control therapies consist of soft multifocal contact lenses and orthokeratology. Soft multifocal contact lenses are typically characterized by a central plano zone for distance vision, with an annular ring of positive refractive add power (akin to a presbyopic multifocal). Orthokeratology, while worn during overnight sleep, re-molds the cornea to correct refractive error. As a byproduct of flattening the central cornea (to correct myopic refractive error), the peripheral corneal curvature is steepened, inducing negative spherical aberration, comparable to center-far soft contact lenses described above. As a result, both of these approaches increase the eye's depth of focus, which makes the point of best-focus ambiguous and subsequently may interfere with accommodation.^26,65,66 Alternatively, we did not observe a significant increase in accommodative lag in the current study, although MP2 and MP3 degraded vision at near with respect to SV. MP1 however decreased accommodative lag, further preventing hyperopic defocus in the fovea. A potential advantage of the designs presented here is their “non-image-forming” characteristics in the near, due to the lateral offset of the aberrations in the peripheral optical zones. As a result, and even if they offer some DOF expansion in eyes without accommodation, as we have shown at least in emmetropes, the lack of a single sharp image at near should deter the eye to relax accommodation and use the peripheral zones at near, or even more accurately accommodate using the central plano region (as it appears to happen with MP1). The decentration of the peripheral aberrations was intended to optimize the peripheral wavefront correction at an eccentricity of 30°. It should also be noted that the four quadrants proposed in this lens design provide the intended correction for four points in the retina. Consequently, large swaths of the retina will not receive the intended therapeutic optical correction. Furthermore, peripherally circularly asymmetric designs, such as those examined here, raise the question of whether they should be ballasted to prevent rotation during wear. 

Although AO visual simulators provide an efficient method for evaluating various optical designs without the need for fabricating physical lenses, there are drawbacks. First, this study evaluated vision monocularly, and accommodation was only driven by monocular blur cues, not binocular fusion. Under binocular viewing conditions, we may expect a further improvement in VA and accommodative response due to binocular summation,^67,68 and vergence driving the accommodative triad, respectively. Second, the present study was performed in monochromatic light, which allowed for the removal of the contraindication of chromatic aberrations, presenting a departure from naturalistic polychromatic vision. Future work will include simulations and accommodation measurements in a binocular see-through visual simulator, where proximity and convergence cues are present, and alternative simulating strategies not subject to chromatic artifacts. 

In conclusion, this study investigated the visual performance of several novel optical designs for myopia control contact lenses. What set these designs apart, is their customization to target the optical quality and anisotropy of the peripheral retina through tailored wavefront aberrations. Our findings indicate that these designs offer favorable visual performance and do not worsen accommodative errors, distinguishing them from other multifocal soft contact lens interventions for myopia control. The design showing the most potential (MP1) stood out for its combination of high VA and enhancement of accommodation. This design was specifically engineered to reorient the peripheral point spread function in the average myopic eye from a circumferential orientation to a radial one, mirroring the pattern observed in emmetropic eyes. 

Acknowledgments

The authors thank Karteek Kunala, Maria Vinas, David Fernandez, and Keith Parkins for technical assistance in the development of the Adaptive Optics Visual Simulator. 

Supported by Empire State Development Funds (New York), Center of Emerging and Innovative Science (University of Rochester) and Clerio Vision, Inc. (Rochester, New York); National Institutes of Health, National Eye Institute NEI R01-EY035009 and P30EY 001319; Unrestricted Funds Research to Prevent Blindness 

Disclosure: S. Aissati, None; T. Zou, None; S. Goswami, None; L. Zheleznyak, Clerio Vision Inc. (P, E); S. Marcos, None