Refractive Focus

Predicting Visual Performance From Optical Performance

Refractive Focus

Predicting Visual Performance From Optical Performance


By Jason Marsack, PHD

Whether as a clinician performing an eye exam or as a patient undergoing the exam, we are intimately familiar with the process of subjective refraction. Patients are presented with a series of two choices, and for each set they choose which one they prefer.

While the choices being made by patients during the exam represent a task that is reasonably straightforward, the process by which patients make those choices is quite complex. Image formation by the optics of the eye, image sampling at the retina, and neural processing in the retina and brain all contribute to the final visual percept that patients “see.” The process of refraction repeatedly provides two separate percepts that patients are asked to judge. The fundamental difference between any two choices is the quality of the retinal image formed by the combined optics of the eye and the trial lens.

Subjective Versus Objective Performance

Refraction is a subjective technique (requiring patient feedback) that quantifies the optical quality of the eye. Tools such as wavefront sensors provide an objective (no feedback needed) measure of the optical quality of the eye. The operating principles of wavefront sensors are conceptually simple: a small spot of light is placed onto the retina, which acts as a source, radiating light backward through the optics of the eye. The optical imperfections in the eye’s optics deviate the light from a reference path as it propagates out of the eye, and these deviations are quantified by the wavefront sensor. The measured deviations (aberration) reported by the wavefront sensor provide a comprehensive description of the optical performance of the eye (including both low-order and high-order aberration) for a single wavelength of light.

Applying Image Quality Metrics to Predict Visual Performance

Once the optical properties of the eye are known, they can be used to understand how images are formed on the retina. This begs the questions: Can knowledge of a patient’s optical quality be used to predict the quality of the resultant visual percept, and by extension, can subjective visual performance be anticipated from objectively measured optical performance?

Researchers and clinicians have always been interested in the impact of optical quality on the visual percept. As one example of the kind of work being conducted in this area, a paper was authored by Thibos et al in 2004 relaying 31 different optical quality metrics (metrics). These metrics take as input the optical performance of the eye measured by a wavefront sensor and provide a single-valued output describing some aspect of the optical quality of the eye. Some of the metrics were commonly known in vision science such as root mean square of the wavefront error, which describes the flatness of the wavefront. However, some were novel formulations such as neural sharpness, a metric described by Williams (2003) that quantifies the effectiveness of a point spread function (PSF) at stimulating the neural visual system.

How are researchers using these metrics? I’ll discuss a brief sample of the findings in this area to demonstrate the type of work being done to relate image quality and visual performance.

Cheng et al (2004) used an experimental paradigm that blurred the target, rather than imposing optical blur on the eye, to demonstrate that certain metrics are well correlated with visual acuity and that these same metrics are good at predicting subjective best focus.

Again using blurred letter targets, The Visual Optics Institute examined the correlation between the metric values and visual acuity in the presence of two experimentally varied high-order aberration modes (Marsack et al, 2004). Several metrics in this study were found to be well correlated.

Chen et al (2005) utilized a deformable mirror to study a variety of image quality conditions. They demonstrated that metrics are capable of predicting the perceived sharpness of images in the presence of aberration.

Interestingly, Shoneveld et al (2009) demonstrated that visual performance in keratoconic eyes can be predicted from metrics. However, they also stated that the predictive metrics in keratoconic eyes are different from those used in predicting performance in normal eyes.

Ravikumar et al (2011) demonstrated that metrics are capable of stratifying six just-noticeable differences in image quality before one line of visual acuity is lost, demonstrating that a perceived change in image quality may occur before a clinically measurable change in visual acuity.

Lastly, Ravikumar et al (2012) demonstrated that change in several of the metrics reported by Thibos are highly correlated with change in visual acuity over a range of different pupil diameters.

As a group, these studies provide a brief snapshot of how investigators are utilizing optical quality metrics to understand patients’ visual performance in the presence of aberration.

How Might Optical Quality Metrics be Employed?

It is true that the predictive metrics of optical quality discussed above remain in the research stage regarding their ability to predict subjective visual performance. In particular, metrics will gain clinical relevance only if they are predictive on an individual basis. If they prove to be so, there are many potential clinical applications of predictive metrics. I’ll briefly discuss two such potential uses as they apply to patients who have high levels of optical aberration, such as patients who have keratoconus.

Predict performance prior to building a custom correction. The feasibility of custom wavefront-guided contact lenses for keratoconus was a topic that I discussed in my article, “Incorporating Wavefront Error Correction in Contact Lenses” that appeared in the September 2012 issue. As the question now is not can custom wavefront-guided contact lenses be built for a patient, but should they be, it becomes useful to have a prediction of the potential improvement associated with the custom correction. The clinician and patient could then weigh the potential gain in visual performance with other factors such as cost, chair time to fit, comfort, care regimens, etc.

Optimize a low-order correction for an optically complex patient. Refraction in normal eyes works well because the sphero-cylindrical aberrations dominate the eye’s refractive error. However, in optically complex eyes exhibiting high levels of high-order aberration, there may be several low-order refractions that provide similar retinal image quality due to interactions between low-order and high-order aberration terms. Predictive metrics may allow identification of optimal sphero-cylindrical corrections from a patient’s objective optical quality data, allowing the clinician to evaluate a host of corrections prior to determining a final prescription.


A large number of physical and physiological aspects factor into the formation of the visual percept. The development of metrics that operate on objective aberration data to predict visual performance, or a change in visual performance, may in the future provide clinicians with yet another tool to assist in understanding the visual percept experienced by their patients. CLS

To obtain references, please visit and click on document #207.

Dr. Marsack completed a PhD in Physiological Optics and Vision Science at The University of Houston, College of Optometry. His research interests include optical aberration of the eye, custom and pseudo-custom correction of optical aberration, visual performance, metrics predictive of visual performance, and ocular drug delivery.