Ying-Ling Chen et al.

studies of tear film models.19−21 These studies show that the optical parameters of the tear film that can affect the accuracy and completeness of optical eye modeling include tear film thickness, post-blink tear undulation, tear breakup pattern, eyelid-produced bumps and ridges, bubbles, and rough pre-contact lens tear surfaces. These tear film characteristics in spatial and temporal domains can be included in the schematic eye models. The predictive modeling and simulation could yield insightful information regarding the dry eye vision and promote the diagnostics technology for the disease.

13.3.1.4. Optical opacity

The only published cataract eye model is proposed by Donnelly.22 In this study, Scheimpflug cameras characterized the anterior segment and backscatter from cataract. He discussed how to measure and model intraocular light scatter with SH wavefront-sensing data.22 The key to simulate the cataract based on our personalized eye models is to simulate the surface and volumetric scattered light in the eye, both of which contribute to contrast reduction of the image at the retina. The scattering theory to be applied is Rayleigh-Mie scattering, since cataracts are volumes composed of small scattering particles.22,23 The scattering of cataracts can be simulated with a bi-directional scatter distribution function (BSDF) in a nonsequential component in ZEMAX.

13.4. Examples of Ophthalmic Simulations

Eye models have applications in scientific research and industrial engineering. Patient vision simulation is useful for medical education and for patient consultation. Ophthalmic measurement simulations are advantageous for even more applications. As demonstrations, the simulations of retinoscopy and the photorefraction (PR) measurement are described in this section.

13.4.1.Simulation of Retinoscopy Measurements with Eye Models

Retinoscopy was introduced more than 100 years ago and is still practiced today to yield important clinical results. A traditional “spot retinoscope”

406

Optical Eye Modeling and Applications

projects a spotlight onto a patient’s eye at a distance of 0.5–1.0 meter, while a contemporary “streak retinoscope” projects a straight-filament image. The size of the spot or streak projection is adjustable by moving a condenser lens that is located above the light source (or filament) in the handle of the device. The retinal reflex is observed by the examiner through a peephole on the scope. When moving the streak projection across the patient’s pupil, the reflex of a myopic or hyperopic eye appears to move with or against the projection motion. The moving speed and direction of reflex depend on the position of the condenser lens. Subsequently, the examiner uses a phoropter or manually places trial lenses over the examinee’s eye to “neutralize” the reflex movement.

When the refraction is neutralized, the pupil will suddenly appear bright, as the light projection aligns to the center of pupil, and will turn dark if the projection slightly misaligns toward either side. No movement should be seen under the neutralization condition. The compensation lens indicates the required defocus correction. Retinoscopic measurement is objective and, therefore, especially useful in prescribing corrective lenses for patients who are unable to undergo a subjective refraction test that requires a response and judgment from the examinee. Retinoscopic measurement is also used to evaluate the accommodative ability of the eye and to detect latent hyperopia. Although the retinoscope optical structure is simple, the thorough analysis is not easy due to frequent utilization of low-cost, imprecise optical elements. As a consequence of the absence of detailed analysis, medical textbooks illustrate retinoscopy with over-simplified portraits. Further, ambiguous observations occur when the ocular aberrations are significant and when overlap occurs in the light path of the multiple aperture stops, including eye pupils. These limitations and difficulties have perhaps discouraged the quantitative of retinoscopy in clinical practice.

With eye modeling, retinoscopic measurements can be predicted through simulation. The optical layout of the simulation is shown in Fig. 13.7. Using a program such as ZEMAX, light rays from the filament source of retinoscope are traced through the scope and enter the targeted eye model. In the return path, the retinal image at the end of the first ray tracing serves as the light-source object, and the light rays are traced through the peephole of the scope and imaged on the retinal plane of the observer. The eye model of the retinoscopist can be a perfect thin lens with an image

407

Ying-Ling Chen et al.

Fig. 13.7. Optical layout in the retinoscopy simulation.

plane. As in the real condition, four effective apertures were involved in this retinoscopic simulation. These apertures were the small apertures in front of the filament, the window on the beam splitter (along both paths), the pupil of the eye (along both paths), and the peephole of the observation. A coordinate break in the ZEMAX program is required to simulate the movement of the streak projection, as it moves across the examinee’s pupil. Ray aiming is necessary to ensure that all of the vignetting or cut-off effects are encountered when using coordinate breaks.

Figure 13.8 demonstrates some example results of the retinoscopy simulations with eye models. On the left of the figure is the famous scissors reflex of a KC eye, as the streak retinoscope projection moves along the 45-degree meridian. Typically, 100 million rays are traced in each of these double-path simulations to produce one reflex image. A KC patient’s topography and manifest refractive prescription are used to construct a personal eye model. Similar simulations can be performed to predict the observation when the eye is focused on a different direction and when the eye pupils are dilated to different sizes. Likewise, the retinoscope simulation can be assigned to move along any meridian with any sleeve position of the scope. In Ref. 24, both planeand concave-mirror practices of retinoscopy are simulated. The typical ammetropia reflex movements of withand against-motion, and the

408

Optical Eye Modeling and Applications

Fig. 13.8. Simulations of retinoscopic observations on (a) a 38-year-old Caucasian female patient’s KC eye and (b) an eye with high degree of SA resulting from a surgical procedure. Scissors reflex and hourglass reflex are illustrated, respectively.

so-called “anomalous with-motion” of the high myopia condition are produced. This type of simulation can be applied to medical training, without the need for using human subjects.

A second retinoscopic example is the hourglass reflex that is shown in Fig. 13.8(b). In the recent years, case observations have suggested that inadvertently induced SAs from surgical procedures, such as the Schachar’s sclera band procedure and the use of intraocular lenses produce “pseudoaccommodation” in presbyopia patient vision. One such case was reported by Dr. Guyton in Johns Hopkins.25,26 Dr. Guyton had the opportunity to examine a patient after surgery for presbyopia. The patient had relied on reading glasses to see objects that were nearby and had undergone Schachar’s scleral band procedures for presbyopia. Two to three months later, she went without glasses entirely, with 20/20 uncorrected VA for viewing both objects in the distance and objects nearby. However, in the dynamic retinoscopic reflex examination, Guyton observed the static hourglass shape of reflex, rather than the streak reflex, for both near and distant visions, thereby demonstrating the absence of actual accommodation. He suspected that the hourglass reflex indicated a condition of a high-degree of SA, which provides an effect of long focal depth, or the so-called “pseudoaccommodatrion.” This hourglass reflex observation can be reproduced with the retinoscopy simulation using general eye models.

409