Chapter 215 – Psychophysical Tests for Glaucoma
ANDREW J. MAYS
• The study of human visual pathway responses to various visual stimuli.
• Subjective tests require a conscious response; these include frequency-doubling perimetry, ring perimetry, acuity perimetry, motion detection perimetry, pattern discrimination perimetry, contrast sensitivity, and color vision.
• Objective tests require no conscious response; these include the electroretinogram and multifocal visual evoked potential.
Psychophysical tests have been developed in addition to traditional perimetry to detect early glaucomatous visual loss. These tests include frequency doubling, high-pass resolution (ring), acuity, motion detection and pattern discrimination types of perimetry, contrast sensitivity (both spatial and temporal), color vision, electroretinography, and visual evoked potential (VEP). If glaucomatous damage is selective to one or more neuropathways, isolating and testing these pathways may permit disease detection before changes appear in the optic disc, nerve fiber layer, and visual field. The proposed visual pathways and the tests designed to stimulate them are described herein.
Visual field testing is imperfect. Clinically, visual field loss often correlates with nerve fiber layer loss and optic nerve damage. However, a substantial portion of the peripheral ganglion cells must be damaged before the loss is detected even by automated perimetry. Also, viable ganglion cells have been found in areas unresponsive to conventional perimetric stimuli. Clearly, automated perimetry does not thoroughly evaluate visual function. Therefore, other tests to enable earliest possible detection of field changes are under development.
Tests designed to evaluate retinal visual function fall into two categories—subjective and objective. The objective tests require no conscious effort on the patient’s part. These tests include electroretinography (ERG) and multifocal VEP. The subjective tests require conscious acknowledgment of the effect of the stimulus. That is, the patient must indicate purposefully whether the stimulus has been perceived. These tests include all those noted previously other than ERG and VEP.
PURPOSE OF THE TESTS
Psychophysical tests are designed to isolate various retinal pathways. In their simplest form, these pathways can be broken down into three sets of cells at different retinal layers. The first cell layer consists of the photoreceptors—rods and three types of cones. Rod function can be isolated through dark adaptation. However, photoreceptor loss has not been proved in glaucoma. Also, retinal pathways, not individual cell types, are of primary interest in psychophysical tests. Scotopic conditions, which isolate rod function, are not used in perimetry. The cones absorb light at different wavelengths and are subdivided into L cones for long (red) wavelength absorption, M cones for medium (green) absorption, and S cones for short (blue) absorption.
The second cell layer is in the middle retina and contains the bipolar cells, which process the information from the photoreceptors into two color-integrating pathways. One pathway carries stimuli from the L and M cones and, because it is concentrated in the fovea, it is the proposed route of spatial contrast information. The other bipolar cell pathway carries information from the S cones (blue wavelength); it constitutes less than 15% of the total cone population and is usually absent from the fovea. This pathway seems to be damaged preferentially in glaucoma, which results in the so-called tritan (blue-yellow) defect.
The third cell layer is defined by two populations of ganglion cells. One population, the tonic or P cells (named for their projection to the parvicellular region of the dorsolateral geniculate nucleus), constitutes about 80% of the ganglion cells. These ganglion cells are found mostly in the fovea and are thought to be responsible for color vision and high-contrast acuity. The P cells are smaller than the phasic or M population of ganglion cells (named for their projection to the magnicellular region of the dorsolateral geniculate nucleus). The M cells constitute about 10% of all ganglion cells and are thought to be responsible for high-frequency temporal-contrast sensitivity, low-frequency spatial-contrast sensitivity, and motion detection. New psychophysical tests and new applications of existing tests are being developed to recognize early glaucomatous damage by examination of these various pathways ( Table 215-1 ).
TABLE 215-1 — VISUAL PATHWAYS AND PSYCHOPHYSICAL TESTS THAT ISOLATE THEM
Retinal Cell Layer
Tests That Isolate Subsets
S cones (blue–yellow pathway)
Short-wave automated perimetry
Cones: L, M, S
Farnsworth–Munsell, Farnsworth, and L’Anthony
L and M cells (spatial contrast)
Ring perimetry, acuity perimetry
Motion detection perimetry, frequency doubling perimetry
UTILITY OF THE TESTS
Several inherent properties of psychophysical tests limit their usefulness. Subjective tests require threshold determination; threshold is defined as a point at which the stimulus is perceived 50% of the time (i.e., the subject responds correctly 50% of the time). Determination of this point allows intra- and intertest comparison, but the point may vary from day to day as a result of a number of factors, which include the patient’s condition and the test conditions. Interpretation of data for subjective and objective tests requires statistical analysis of percent sensitivity and specificity. Sensitivity refers to the ability of the test to detect glaucoma, and specificity refers to the ability to detect the absence of glaucoma. The results are commonly reported through the use of receiver operating characteristic (ROC) curves in which the percent of 1 – specificity, equation (215-1) , is plotted against the percent sensitivity ( Fig. 215-1 ).
A given psychophysical test, over a range of selected values, generates an ROC curve, and a trade-off always exists between sensitivity and specificity. If sensitivity is increased, patients who have the disease could be excluded. If sensitivity is decreased, normal individuals could be included in the disease group. After the data have been plotted on an ROC curve, an optimal value to separate patients with glaucoma from normal subjects is selected. The ROC curve may be critical in deciding whether a particular test is sensitive and specific enough to use in the general population.
Glaucomatous visual field loss is often detected later in the disease; evaluation of the optic nerve head is a better way to detect early disease progression. Shortcomings with psychophysical tests designed to detect early field loss stem from external and internal sources. Externally, the test depends on the patient. Patient understanding and cooperation, patient age, pupil diameter, media opacities, learning effects, and so forth affect test results. Internally, the test relies on statistical analysis and probability. Data generated from an individual are compared with data generated from a large, demographically
Figure 215-1 Receiver operating characteristic curves. Curves are shown for three arbitrary levels of separability between patients with glaucoma and normal subjects. Each curve or line is generated by a hypothetical test over a range of selected thresholds. Separability is related directly to the relative area under the curve or line. Curve a, with an area of 50% above and below, is an invalid test as its results are no better than chance. Curve b shows greater area under the curve but does not approach usefulness for the detection of glaucoma. (Remember that the test must detect a disease found in 2–3% of the general population.)
similar population. The probability of the presence of glaucoma is determined, but this information must be used with discretion. Other disease states such as cataract, macular degeneration, and ischemic optic neuropathy may confuse results. Internal and external factors must be kept in mind when psychophysical test results are interpreted. At present, visual field tests must be used adjunctly to determine the progression of glaucomatous damage.
The ideal new test should have good sensitivity and specificity and be user friendly. Several new tests show promise.
A recent application of contrast sensitivity combines spatial and temporal contrast. As mentioned earlier, the M cells seem to mediate low-frequency, spatial-contrast sensitivity and high-frequency, temporal-contrast sensitivity. When a low spatial frequency pattern undergoes high temporal frequency flicker (the bars alternate rapidly between dark and light), the pattern perceived appears to have double the spatial frequency. This phenomenon, described in the mid-1960s, is referred to as frequency doubling. Its application in glaucoma is promising. Targets that consist of rapidly alternating vertical bars are projected onto the central 30–35° field. The patient responds when the vertical bars are seen. Early results correlate well with preexisting visual field loss defined by automated perimetry,  and visual blurring has little effect on the test results. Refinements in the technique are being developed.
High-Pass Resolution Perimetry
High-pass resolution perimetry or ring perimetry is an alternative form of peripheral visual field testing. The test involves a ring of fixed luminance but variable size ( Fig. 215-2 ) that is projected onto 50 locations within the central 30° visual field. As the ring size is adjusted, the threshold is determined. This system of perimetry is thought to stimulate the P cell pathway. Advantages over conventional static perimetry include increased speed, increased patient acceptance, reduced variability, and reduced
Figure 215-2 High-pass resolution (ring) perimetry stimulus. The size of the ring varies but luminance is held constant.
learning curve. High-pass resolution perimetry detects glaucomatous visual field loss as well as static perimetry and may be used more in the clinical setting.
Acuity perimetry is another test thought to detect early glaucoma. A laser interferometer is used to project interference patterns (straight lines) on the retina. The perimeter projects the images on the central 20° of vision along any meridian at 1-minute intervals. The apparatus is the same as that used to assess central visual acuity in patients affected by media opacities. Early studies have shown that acuity perimetry is more sensitive than conventional perimetry in the detection of glaucomatous damage. Two hypotheses support this contention. First, visual acuity is a more complex function compared with the detection of stationary white lights. Second, fewer ganglion cells need to be damaged for a more complex visual function to be affected. Despite these early results, acuity perimetry in this form is not widely available in the clinical setting. Theoretically, ring perimetry and acuity perimetry test similar visual functions.
Motion Detection Perimetry
Motion detection may be related to temporal contrast sensitivity—both seem to be modulated primarily by the M ganglion cells. In administering the test for motion detection perimetry, the patient observes many dots on a screen. Subsets of dots are moved coherently against a random background of dots ( Fig. 215-3 ). The threshold is calculated by determination of the number of moving dots necessary to detect motion, and this number is divided by the total number of dots. Motion detection is decreased in glaucoma patients, but further refinement in the technique is necessary before it is clinically applicable.
Figure 215-3 Motion detection perimetry stimulus. 0% coherence: all the dots move randomly; 25% coherence: one quarter of the dots move in the same direction; 50% coherence: one half of the dots move in the same direction.
Pattern Discrimination Perimetry
Pattern discrimination perimetry involves a checkerboard image superimposed on a random dot background (the image is projected onto 32 points within the central 30° of vision). Deliberate changes in the proportion of random dots to checkerboard dots make the stimulus more difficult to perceive ( Fig. 215-4 ). Pattern discrimination is a more complex visual function than on-off white light recognition. In theory, as with ring and acuity perimetry, the recognition of pattern requires a greater number of intact ganglion cells, and damage should be detected earlier if this theory is correct. Preliminary studies show a decrease in pattern discrimination function in glaucomatous eyes. Visual loss detected by pattern discrimination perimetry does not seem to correlate with nerve fiber loss and optic nerve changes, which may reflect damage to the visual system undetectable by direct observation of the optic nerve and retinal nerve fiber layer. Continuing work with this perimeter will determine its future role in glaucoma.
Contrast sensitivity measures luminance between brightly and dimly lit areas, and spatial contrast uses adjacent areas of light and dark. The distance between the areas (frequency) and the intensity of the areas (luminance) are adjusted over a range. The highest possible spatial contrast sensitivity corresponds to the greatest visual acuity ( Fig. 215-5 ). Temporal contrast uses one area that flickers between light and dark. Both spatial and temporal contrast sensitivities are decreased in glaucoma patients. Spatial contrast sensitivity is nonspecific and unlikely to be useful in the detection of early glaucoma. The usefulness of temporal contrast sensitivity is unclear. The majority of the current reports in the literature suggest that temporal contrast sensitivity is decreased early in the course of glaucoma. However, the clinical usefulness of this information remains under investigation.
Color vision is affected by glaucoma. Most reports in the literature support the preferential loss of S cones (blue wavelength) as a result of glaucoma, but as yet no explanation exists for this observation. As S cones are absent from the fovea, short-wave (blue-on-yellow) perimetry, discussed in Chapter 214 , is used increasingly on this basis. In the absence of perimetric testing, three common color vision tests, the Farnsworth-Munsell 100-hue test and the Farnsworth D-15 and L’Anthony D-15 desaturated color tests, may be employed to evaluate glaucomatous damage. All detect a tritan defect equally well. The simplest color test is the most practical in a clinical setting. Other conditions, such as cataracts, macular degeneration, and multiple sclerosis–associated optic atrophy, may cause tritan defects. However, color vision tests do not provide useful clinical information in the diagnosis of early glaucoma. These tests are unlikely to supplant currently employed tests to assess visual damage as a result of glaucoma.
Figure 215-4 Pattern discrimination perimetry stimulus. 0% coherence equals random checker pattern in center. 100% coherence equals perfect checkerboard pattern in center. The computer generates patterns between 0% and 100% coherence.
Figure 215-5 Spatial contrast sensitivity. A, Low contrast. B, High contrast. C, Snellen acuity. The frequency is held constant, but the luminance of the bright stripes is increased and the luminance of the dark stripes is decreased (note relationship to visual acuity).
Electrophysiologic tests are objective. Patient interaction is minimal, and the tests are minimally invasive. ERG has been applied to glaucoma patients using both flash and pattern techniques. Flash ERG stimulates photoreceptors (a-wave) and bipolar and Müller cells (b-wave). Ganglion cells contribute little to the recorded response, and, therefore, the flash electroretinogram is not affected appreciably even in advanced glaucoma. Pattern ERG averages responses to an alternating checkerboard, bar, or sine pattern. Part of the pattern ERG response may originate from the ganglion cells, but the source of the pattern response is not understood fully. Evidence has accumulated that suggests that pattern ERG amplitudes are decreased in glaucoma.  In particular, the negative wave measured at 95?ms (N95) decreases significantly in glaucoma ( Fig. 215-6 ). 
Multifocal Pattern Visual Evoked Potential
Visual evoked potential (VEP) detects cerebral cortical responses to visual stimulation. Glaucoma associated with visual field loss causes an abnormal VEP. By applying electrodes straddling the inion, subtle visual field defects within 26° of fixation can be detected. Although in its early stages of development, this technique shows promise as an accurate and objective measurement of glaucomatous visual loss.
All newer tests for perimetry are administered in much the same way as automated perimetry. A central fixation target is used,
Figure 215-6 Pattern electroretinogram, normal trace. P50 is the positive deflection at 50?ms. N95 is the negative deflection at 95?ms. In glaucomatous patients, all pattern ERG parameters, especially those at 50 and 95?ms, are decreased in amplitude. Normal eye: P50 = 2.97?mV ± 0.77?mV; N95 = 1.39?mV + 0.62?mV. Glaucomatous eye: P50 = 1.43?mV ± 0.47?mV; N95 = 0.13 + 0.16?mV.
and stimuli are presented in the periphery. To take the test, the patient must be able to sit comfortably for several minutes, be able to respond to the stimulus, and understand the nature of the test. Video monitors allow observation of the patient’s ability to maintain fixation. The patient is asked to push a button when a stimulus is perceived, and computer-generated results (in various forms) are printed. No part of these tests is invasive.
Color testing, another form of noninvasive testing, is simple. Discs of various hues are placed randomly before the patient, who must arrange the colors from darkest to lightest. The results are compared against predetermined patterns of color loss. Less understanding is required of the patient in these tests than in perimetry testing. Still, the patient must be able to sit at a table, move the colored discs, and appreciate the nature of the test.
ERG is slightly more invasive than the other tests. A gold foil, corneal contact electrode is placed on the anesthetized eye and a reference electrode is placed on the skin near the lateral canthus. The patient sits in front of a bowl (similar to that of a Goldmann perimeter) that diffuses light over the entire retina during stimulation. Both eyes may be tested simultaneously. The stimulus is presented and the results are recorded on a computer-generated printout for interpretation. No conscious effort is required of the patient.
INTERPRETATION OF RESULTS
All psychophysical tests are imperfect, and no individual test allows a diagnosis of glaucoma. Low specificity and sensitivity limit clinical application. The subjective nature of many of the tests limits the validity of the results. Therefore, the current, single, best way to identify and follow visual loss from glaucoma is through static threshold perimetry. This type of perimetry has the largest database for comparison, and familiarity with the reported results allows more confidence in interpretation. Most other psychophysical tests can detect visual loss from glaucoma. Ring perimetry and frequency-doubling perimetry may allow the earliest detection of visual loss. Greater experience using the tests and an expanding database will allow better clinical usefulness. As understanding of the complex visual system increases, new tests and new applications of existing tests may enable visual function to be evaluated more completely.
Also, combinations of newer tests are being assessed. In theory, several different visual pathways are tested and the results are combined to give an overall picture of glaucomatous damage. This remains experimental but shows promise in the detection of early glaucomatous field loss. The ideal goal is to detect glaucoma before it causes any visual loss.
1. Quigley HA, Dunkelberger GR, Green WR. Retinal ganglion cell atrophy correlated with automated perimetry in human eyes with glaucoma. Am J Ophthalmol. 1989;107:453–64.
2. Kelly DH. Frequency doubling in visual responses. J Opt Soc Am A. 1966;56: 1628–33.
3. Johnson CA, Samuels SJ. Screening for glaucomatous visual field loss with frequency-doubling perimetry. Invest Ophthalmol Vis Sci. 1997;38:413–25.
4. Graham SL, Drance SM, Chauhan BC, et al. Comparison of psychophysical and electrophysiological testing in early glaucoma. Invest Ophthalmol Vis Sci. 1996;37:2651–62.
5. Phelps CD. Acuity perimetry and glaucoma. Trans Am Ophthalmol Soc. 1984; 82:753–91.
6. DeYoe EA, VanEssen DC. Concurrent processing streams in monkey visual cortex. Trends Neurosci. 1988;11:219–26.
7. Silverman SE, Trick GL, Hart WM. Motion perception is abnormal in primary open-angle glaucoma and ocular hypertension. Invest Ophthalmol Vis Sci. 1990;31:722–9.
8. Nutaitis MJ, Stewart WC, Kelly DM, et al. Pattern discrimination perimetry in patients with glaucoma and ocular hypertension. Am J Ophthalmol. 1992;114:297–301.
9. Chauhan BC, LeBlanc RP, McCormick TA, et al. Correlation between the optic disc and results obtained with conventional, high-pass and pattern discrimination perimetry in glaucoma. Can J Ophthalmol. 1993;28:312–16.
10. Stamper RL, Lerner LE. Psychophysical techniques in glaucoma. In: Ritch R, Shields MB, Krupin T, eds. The glaucomas, Vol 1. St Louis: Mosby–Year Book; 1996;701–13.
11. Bassi CJ, Galanis JC, Hoffman J. Comparison of the Farnsworth-Munsell 100-hue, the Farnsworth D-15, and the L’Anthony D-15 desaturated color tests. Arch Ophthalmol. 1993;111:639–41.
12. Breton ME, Drum BA. Functional testing in glaucoma: visual psychophysics and electrophysiology. In: Ritch R, Shields MB, Krupin T, eds. The glaucomas, Vol 1. St Louis: Mosby–Year Book; 1996;677–99.
13. O’Donaghue E, Arden GB, O’Sullivan F, et al. The pattern electroretinogram in glaucoma and ocular hypertension. Br J Ophthalmol. 1992;76:387–94.
14. Klistorner A, Graham SL. Objective perimetry in glaucoma. Ophthalmology. 2000;107:2283–99.