These two equations, used to calculate the contrast, are the basis of all clinical tests for the evaluation of contrast sensitivity. The contrast is a physically measurable magnitude that corresponds to the information coming from the stimulus. When we present a variable contrast stimulus to an observer, decreasing its magnitude, there will come a time that it cannot be detected. We say then that we have reached the Contrast Threshold.

The threshold of contrast can be defined as the minimum amount of contrast needed to have the stimulus on a uniform background to be detected. If the contrast is lower than our threshold value, we will not be able to detect it. Then appears an ability of the observer that is related to its ability to detect an object on a background. This ability is called Contrast Sensitivity (CS). The contrast (physical magnitude, C) perceptible by an observer will reach a threshold value whose inverse will give rise to contrast sensitivity (visual ability, CS = 1/C).

HOW DO WE MEASURE CONTRAST SENSITIVITY?

We already talked in the previous section that we can use sinusoidal or variable contrast optotypes to assess contrast sensitivity. However, the differences between the two visual stimuli mean that the information obtained with each of them is not interchangeable, although it is related in some way3,4.

Sinusoidal networks are the simplest stimulus pattern for measuring contrast sensitivity. As we can see in Figure 1B there is a degraded transition between each maximum and minimum luminance. This combination is called a cycle and will be used to describe the contrast sensitivity test. The more cycles there are within a degree we will say that the network has a greater spatial frequency and vice versa. In contrast sensitivity tests composed of sinusoidal networks, the so-called Contrast Sensitivity Function (CSF) is analyzed. To determine this curve, the sensitivity (1/C) is calculated with the increase of the spatial frequency. The result is a curve where the evaluated spatial frequencies are represented on the abscissa axis and the sensitivity for each of these frequencies on the ordered axis.

The sensitivity obtained by increasing the frequency of a sinusoidal network is correlated with the sensitivity obtained by decreasing the size of an optotype. Although it is true that this correlation is not linear5, the sensitivity measured with high frequency networks maintains a better agreement with the sensitivity measured with optotypes of small size or less detail and vice versa. Figure 2 describes an analogy between the CSF measured with sinusoidal networks and the agreement with optotypes of variable size. The curve would represent the CSF in such a way that contrasts above the curve will not be perceived by the patient as opposed to contrasts below the curve. We also see that as the frequency increases and the size of the optotype decreases so does the contrast sensitivity in the normal patient. A metric related to this curve would be the visual or inverse acuity of the smallest optotype detail that a patient is able to solve.

Both high- and low-contrast visual acuity would represent only one point of the curve, which shows the loss of information about total visual performance when we only measure visual acuity and not CSF.

The CSF measurement tests in multifocal intraocular lens studies do not measure the CSF completely but select 4 or 5 spatial frequencies. For example, the VCTS6,7 and the FACT8,9, two of the most used CSF measurement tests incorporate 5 spatial frequencies (1.5, 3, 6, 12 and 18), with the separation between frequencies being one octave. The CSV1000-E and the ClinicCSF10, on the other hand, only evaluate 4 channels (3, 6, 12 and 18) separated by an octave11 and inclining the sinusoidal networks ±15º to subject the patient to a multiple choice task with 3 possible answers (vertical, tilted to the right and tilted to the left)2.

To expedite this procedure and minimize possible errors in the distance of presentation we can use a remote distance test (~ 6 m) and introduce different positive and negative lenses that vary the vergence of light in the same way as if we varied the distance of presentation of the test. To calculate the correspondence between distance and diopters we just have to calculate the inverse of the distance in meters. For example, to simulate a distance of 40 cm we would have to place a power lens equal to -2.5 D (1/-0.4 m).

The correlation between dioptric jumps and distance is not linear as shown in Figure 3. This means that a small variation in the near distance presentation may correspond to an important dioptric jump. As can be seen in the enlarged region of near in Figure 3. A variation in the presentation distance of 25 cm to 30 cm corresponds to a dioptric jump of more than 0.5 D from a blur lens of -4.00 D to a higher than -3.50 D while a dioptric jump of 0.50 D in far distance, for example from -1.0 D to -0.5 D, corresponds to a distance variation of 100 cm. Due to this important variation of visual acuity with a small distance error, it is advisable not to measure the near-distance blur curves and always do it at a distant distance18. We already know that it is recommended to measure the blur curves for distance, the greater the distance the better, and if we have no possibility of having a cabinet with a distance of 4 meters or more we can carry out the procedure at 2 m as long as we place a +0.50 D lens during the test that corrects the proximal distance.

The next question we can ask ourselves is, what blur lenses should I use? The answer to this question will depend on the range of vision we want to evaluate or in other words, the working distance for close of the patient. Bearing in mind this consideration, it is true that in a range of +1.00 D to -4.00 D we cover the patient's vision from infinite to 25 cm, as shown in Figure 3. In particular cases this range may vary, for example with intraocular lenses of extended depth of focus (+1.50 D to -2.50 D)19. In addition to the range of vision we must also decide the dioptric jumps that we are going to select; in a standardized way we will always carry out the curves in steps of 0.50 D20.

When we carry out the procedure with optotype charts in patients with intraocular lenses (pseudophakic) it is preferable to start the procedure with negative lenses (-4.00 D) since we will go from a worse vision to a better vision, and in this way we will avoid the problem of memorization of the letters in case we do not have a computerized test in which the optotypes vary randomly with each patient’s response. In addition, in bifocal intraocular lenses in which visual acuity decreases in intermediate vision it would be advisable to change the static optotype chart from -2.0 D in order to avoid memorizing the letters with the consequent over-estimation in intermediate vision.

Another alternative proposed by Gupta et al to avoid the learning effect is to randomize the reading order of the letters, randomize the order of the defocus lenses or both21. This last criterion of randomizing the defocus lens is the one currently used during many clinical trials with optotype charts. However, the aforementioned drawbacks disappear when we use a random optotype test such as the Multifocal Lens Analyzer22.

CONTRAST SENSITIVITY DEFOCUS CURVES

Blur curves of visual acuity are widely used in clinical practice. However, as we have already seen, visual acuity only represents one point in the CSF and patients with the same visual acuity can have a very different visual quality that can be differentiated through a contrast sensitivity test23. In 2016, we started the investigations with a new automated tool designed and programmed by Rodríguez-Vallejo M. for measuring the blur curves that allows not only to evaluate the visual acuity blur curve but also the contrast sensitivity blur curve from the contrast variation of a static-sized optotype22. This tool called Multifocal Lens Analyzer consists of an automated test reproduced through an iPad which has proven to be a useful and capable instrument for measuring visual performance10,24–26.

The system has important advantages with respect to conventional tests for measuring defocus curves, including:

• The measurement of the threshold of vision is independent of the clinician and is achieved through a psychophysical method controlled through the application.
• It uses an optotype (E of Snellen with crowding effect) that is presented in 4 possible orientations in a random way in order to avoid memorization.
• The procedure is repeated for each of the possible defocus lenses from +1.00 D to -4.00 D in steps of -0.50 D, finally obtaining a defocusing curve that describes the vision from infinity to 25 cm.
• The procedure is completed in a period between 7 and 8 minutes. Considerably less time than the conventional measure of defocus curves.

Since the procedure is fully automated, the clinician only has to click on the button that matches the response given by the subject, with the application controlling whether to increase or decrease the size/contrast of the optotype, move to the next Defocus level through a lens shift alert message, or end the test. The measurement protocol with the Multifocal Lens Analyzer involves the following procedure:

• Patient sitting 2 m from the iPad with a + 0.50D lens on test glasses that compensates for the proximal vergence.
• The iPad will remain attached through a tripod centered at the height of the visual axis of the patient with a slight inclination in the case of any kind of external light reflection on the screen.
• Disable the automatic brightness option of the iPad settings section and set an 85% brightness percentage within the application.
• Place the best correction obtained for the patient during the subjective examination.
• Perform the monocular or binocular test according to our needs.
• Explain the test to the patient in order to avoid fatigue during the test.

CLINICAL CASE

The following is an example of a clinical case in which the contrast sensitivity blur curves more accurately represent the patient's vision quality than the visual acuity blur curves. Figure 4 depicts a blur curve of monocular visual acuity in a patient implanted with a low-addition trifocal intraocular lens (+3.00 D in the IOL plane, +2.00 D in the glasses plane).

Despite being taken with the best subjective correction we can verify that the visual acuity for far (0 D) is 0 logMAR (20/20). In addition, the visual acuity is maintained at 0.1 logMAR (20/25) to the blur lens of -4.00 D, which means a visual acuity of 20/25 up to a distance of 25 cm, something really strange in a low-addition trifocal lens. We could say that the visual acuity blur curve is really good, but does that mean that the quality of vision is as good as the visual acuity blur curve represents?

Figure 5 demonstrates that there is an anomaly whereby a flatter visual acuity blur curve is obtained. This anomaly consists of a decentralization and inclination of the lens, that induces an increase in the depth of focus, such that a good visual acuity is achieved higher than expected in intermediate vision and beyond -2.0 D.

However, this improvement in visual acuity does not imply an improvement in visual quality as shown in the contrast sensitivity blur curve of Figure 6.