Contrast Sensitivity and the Contrast Sensitivity Function
Another important dimension in measuring visual function is contrast sensitivity—the sensitivity of an observer to differences in luminance between an object and the background. In general, the higher the contrast, the easier an optotype is to decipher. Over a broad range, the visual system is relatively insensitive to the absolute brightness of a visual stimulus, but is much more attuned to the contrast between adjacent surfaces. For example, the dark ink on a printed page reflects about 10% of the incident light. In comparison, the white paper background has a reflectance of perhaps 90%, regardless of the level of absolute Illumination. Thus, when reading under bright sunlight, we still appreciate the printed text as black, even though the absolute brightness of the reflected light is greater than that reflected from white paper in dim illumination, as in twilight. If the brightness of an object (Imin) and the brightness of its background (Imax) are known, the following formula can be used to measure the degree of contrast between the object and its background:
Thus, for typical printed matter, the contrast is about 80% (90% − 10%)/(90% + 10%). Snellen visual acuity is commonly tested with targets, either illuminated or projected charts, that approximate 100% contrast. Therefore, when we measure Snellen visual acuity, we are measuring, at approximately 100% contrast, the smallest optotype that the visual system can recognize. In everyday life, however, 100% contrast is rarely encountered, and most visual tasks must be performed in lower-contrast conditions.
To take contrast sensitivity into account when measuring visual function, we can use the modulation transfer function (MTF). Consider a target in which the light intensity varies from some peak value to zero in a sinusoidal fashion. The contrast is 100%, but instead of looking like a bar graph, it looks like a bar graph with softened edges. The number of light bands per unit length or per unit angle is called the spatial frequency and is closely related to Snellen acuity. For example, the 20/20 E optotype is composed of bands of light and dark, where each band is 1 arcmin. Thus, for a target at 100% contrast, 20/20 Snellen acuity corresponds roughly to 30 cycles per degree of resolution when expressed in spatial frequency notation. The relationship between spatial frequency and the contrast sensitivity at each spatial frequency constitutes the MTF.
In clinical practice, the ophthalmologist presents a patient with targets of various spatial frequencies and peak contrasts. A plot is then made of the minimum resolvable contrast target that can be seen for each spatial frequency. The minimum resolvable contrast is the contrast threshold. The reciprocal of the contrast threshold is defined as the contrast sensitivity, and the manner in which contrast sensitivity changes as a function of the spatial frequency of the targets is called the contrast sensitivity function (CSF) (Fig 3-7). Figure 3-8 shows a typical contrast sensitivity curve obtained with sinusoidal gratings. Contrast sensitivity can also be tested with optotypes of variable contrast (eg, the Pelli-Robson or Regan charts), which may be easier for patients to use. It is important to perform contrast sensitivity testing with the best possible optical correction in place. In addition, luminance must be kept constant when CSF is tested, because mean luminance affects the shape of the normal CSF. In low luminance, the low spatial frequency fall-off disappears and the peak shifts toward the lower frequencies. In brighter light, there is little change in the shape of the normal CSF through a range of luminance for the higher spatial frequencies. Generally, contrast sensitivity is measured at normal room illumination, which is approximately 30–70 lux.
Various physiologic and pathologic conditions of the eye affect contrast sensitivity. Any corneal pathology that causes distortion or edema can affect contrast sensitivity. Lens changes, particularly incipient cataracts, may significantly decrease CSF, even with a normal Snellen visual acuity. Retinal pathology may affect contrast sensitivity more (as with retinitis pigmentosa or central serous retinopathy) or less (certain macular degenerations) than it does Snellen visual acuity. Glaucoma may produce a significant loss in the midrange of spatial frequencies. Optic neuritis may also be associated with a notch-type pattern of sensitivity loss. Amblyopia is associated with a generalized attenuation of the curve. Pupil size also has an effect on contrast sensitivity. With miotic pupils, diffraction reduces contrast sensitivity; with large pupils, optical aberrations may interfere with performance.
Figure 3-7 Campbell-Robson contrast sensitivity grating. In this example, the contrast diminishes from bottom to top, and the spatial frequency of the pattern increases from left to right. The pattern appears to have a hump in the middle at the frequencies for which the human eye is most sensitive to contrasts.
(Courtesy of Brian Wandell, PhD.)
Figure 3-8 A typical contrast sensitivity curve is noted as x-x-x. The shaded area represents the range of normal values for 90% of the population. Expected deviations from normal due to specific diagnoses are noted in the text discussion.
(Developed by Arthur P. Ginsburg, PhD. Courtesy of Stereo Optical Company, Inc, Chicago.)
Impairments in contrast may be disqualifying in certain occupational situations, such as driving heavy vehicles. Recognition of these difficulties is often valuable in understanding the complaints of patients, who may have difficulty with certain visual tasks notwithstanding good Snellen acuity as measured with the standard high-contrast eyecharts.
Excerpted from BCSC 2020-2021 series : Section 3 - Clinical Optics. For more information and to purchase the entire series, please visit https://www.aao.org/bcsc.